Tern Wholesale · Imani Reed → Nia HoltInitial position: start 4 October 2027 · synthetic planning case

Buy the supported November recovery.

Recommend S + K now for £16,000 and replace the blanket 25 October promise with a gated 8 November target. That supports all three supplied states at £128,000 gross, protects £12,000 hypercare and leaves £22,000 of authority. 17-A17-E–G

The authorisation Nia should give

Approve the date change, stop optional rebates/dashboard development, and purchase S (£6,000) plus conditional K (£10,000) before D1 assignments. Imani can authorise £16,000 within her £25,000 purchase limit; only Nia can change the promised date and ultimately approve cutover. K is paid now and used only if Q verifies the native fault by the end of D17. 17-A17-C17-E

The earliest-window alternative remains executable: buy £32,000 now to aim for 25 October in J1, with 8 November in J2/J3. It leaves just £6,000 in its fallback states. The delivery view favours its expected economics by £12,600; the finance view favours the lower-exposure policy by £400. With no agreed risk preference, the £22,000 uniform funding margin and one supported planning date justify the lower-exposure recommendation. This is a judgement, not dominance under both views. 17-F–G

Planning proposal, not approval or achieved readiness. All schedules below assume the stated work and successful final evidence gates. These three bounded states do not guarantee the real future. No purchases, signatures, accepted records or passed tests have been recorded by this page.

Readiness is materially below “90%”.

65%

Distinct records accepted

9,000 imported rows − 1,200 duplicates = 7,800 distinct. Finance accepts 6,500; 1,300 are pending and 2,200 required records are absent. 3,500 still need acceptance. Supplier “7,000 approved” contains 500 provisionally reviewed pending records. 17-B

12 / 30

Pass on current candidate

Eight fail and ten are unrun. All 30 must be rerun in final UAT, so current passes are no substitute for final acceptance. Six of 20 recent closures reopened (30%); this is a warning about quality, not a state probability. 17-B17-D

60

Closed core tickets

90 closed − 30 optional = 60 core closed. With 0–10 core tickets among the unknown open mix, core closure is 60/(60 + x), or 85.7–100%. Ticket count cannot establish integrated readiness or predict J1/J2/J3. 17-B

Free the core path: defer five E-days of optional rebates and three D-days of dashboards. They collide directly with W1 core assignments shown below. Use manual rebates for at most eight weeks only if approved; do not treat that permission as already granted. The policies retain native N and do not incur X’s four finance hours/week (32 hours over eight weeks). 17-A17-C–D

Compare the decisions available now

Selections are an ephemeral planning view. Reset or reload restores the initial lower-exposure / J1 / delivery-view position. Original evidence remains below.

All state funding gates stay visible

Gross cash is tested state by state; expected values cannot fund a branch.
Policy / joint stateCutoverBase burnUpfrontHypercareGrossOutside valueAuthority left

£4,000 × pre-cutover weekdays + non-refundable purchases + protected £12,000. D16 uses 15 burn days; D26 uses 25; D36 uses 35. Legacy operation is supported on this basis. Trading value is outside the project account. 17-A17-G

Expected comparison

Let p = P(J1); J2 and J3 have the same cash outcome within each policy. Earliest gross = £144,000 − £40,000p; lower-exposure gross = £128,000. Earliest outside trading value = £12,000p; lower-exposure value = £0. Earliest economic cash less value = £144,000 − £52,000p. Economic switch: p = 16/52 = 30.769%; gross-cash switch: p = 40%.

“Cash less value” is only an economic comparison, never the funding request. Above 30.769% J1, earliest wins this limited expected-economic metric; below it, lower exposure wins. No value for other delays or improvement capacity was supplied. Neither view is silently averaged. 17-G

Commit people, then branch on evidence

A Tax/creditB1–B4 Ordered data blocksC Historical ordersN Native interfaceL ConversionU Final UATFix Native repairR Rehearsal
Workday / dateESecond EDSecond DQFObservation / capacity

Each occupied cell is one full person-day; idle days still burn £4,000. Dates follow weekdays from Monday 4 October, no holidays. F is one person even on released days. Bookings and this mechanical feasibility check are forecasts, not acceptance evidence. 17-A17-C–F

Keep the tempting failures visible

Claim / changeWhat survives scrutinyFailure or boundary
Supplier “£92,000 accelerated”£60,000 D16 burn + £32,000 purchases reproduces the quote. It can support D16 in J1 using the earliest schedule above.True gross is £104,000 including hypercare; J2/J3 need £144,000. The quote alone neither promises D16 nor budgets every outcome. 17-H
Delete K from S + K£6,000 purchases; J1/J2 can still reach D26 at £118,000.J3 is unfunded. Fault at end D17; repair D20–22 (F22), restarted U24/27/29, R31–32 → D36. £140,000 + £6,000 + £12,000 = £158,000; £8,000 beyond authority. Switching to X18–19, validating F22 yields the same late U sequence. 17-C–F
Add K to the earliest basketPurchases rise from £32,000 to £42,000; J1 still has D16.J2 remains D26 and costs £154,000. Its tax review cannot activate K; this “extra insurance” breaches authority by £4,000. 17-E–F
Wait to buy after the tax reviewD6 information can change assignments and the planned cutover.Capacity cannot be purchased then. All offers must be bought before D1. The decision tree uses only already purchased capacity. 17-E17-H
Subtract £12,000 from the funding requestIt is a confirmed conditional trading contribution only at D16.It is not project cash. Hypercare stays £12,000, and gross authority stays £150,000 in every state. 17-A17-G–H

D36 alone costs £152,000 before any purchases (£140,000 burn + £12,000 care). It is an operationally supported window, but no positive-purchase D36 branch is funded here. Cancellation, an unpaid pause, later money, unsupported cutover and early K activation are not fallbacks. 17-A

Promise the process and funded options; gate the launch

Decision and reversal points

  • Before D1: Nia chooses the date posture; purchases are fixed and non-refundable. Earliest’s £32,000 needs Nia’s purchase approval. Lower’s £16,000 fits Imani’s limit, but Nia must still approve the date change. 17-A/E
  • End D6: finance independently distinguishes J2. Under earliest, release the D16 promise in J2 and use the D26 witness. Under lower, resequence F and A while keeping D26. J1 and J3 remain indistinguishable. 17-F
  • Second actual U day: earliest detects J3 at end D12; lower at end D17. Stop that U sequence, perform the repair, restart all three U days. Lower activates already reserved K only after Q’s verified D17 report. 17-E–F
  • Before the cutover start: if a gate fails, do not launch. Earliest may use D26 only if a revised feasible sequence still fits its £144,000 funded branch. Lower has no funded D36 option. A new outcome beyond the model requires immediate sponsor re-planning inside existing authority; it is not covered by an invented fallback. 17-A/D/F

Evidence Nia must see

  • Finance F: 10,000 distinct records accepted, signed tax/credit logic and balanced rehearsal reconciliation.
  • Q: all 30 final UAT scenarios passed on the unchanged candidate; after J3, a completely restarted final UAT set.
  • Operations: trained readiness, demonstrated rollback and signed operational acceptance from rehearsal.
  • Imani → Nia: named evidence pack, funded window and explicit cutover recommendation; Nia authorises the launch.

No configuration or data changes during a U sequence. A known fault terminates that sequence before repair and full restart. L does not repeat after J3 because converted data is unaffected. 17-D17-F

Responsible external position: “We propose an 8 November all-depot launch, subject to final finance, test, rollback and operations acceptance. Recovery capacity and cash cover all three supplied planning states.” The earliest alternative could promise only a conditional 25 October opportunity, with 8 November as its funded scenario fallback. No unconditional launch guarantee is supportable.

Original supplied evidence

Read-only source records. Calculations, scheduling choices and the recommendation above are derived proposals. No external facts, probabilities or approvals were added.

17-A — Sponsor mandate, Nia Holt, start 4 October 2027.
Sponsor mandate, Nia Holt, start 4 October 2027. D1 is Monday 4 October; count weekdays, no holidays. A task finishes at its last day’s end; successors start next workday. Cutover is at the START of D16=25 October, D26=8 November or D36=22 November only. All three depots launch together; ordering, inventory, purchasing, tax and reconciliation remain core. Rebates/dashboards may be deferred; approved manual rebates may operate eight weeks. Choose upfront purchases before D1 assignments begin; their supplied D1 capacity is then usable. No state information is available at that decision.

Total remaining authority is £150,000, including an immutable £12,000 post-launch hypercare provision for any cutover date. Do not spend or net off that provision. Base programme burn is £4,000 per working day from D1 through the day before cutover, including idle days and all base people/vendors. Purchases are additional. Legacy operation through D36 is supported within this basis; no unpriced pause/discount is available. Imani may authorise aggregate recovery purchases≤£25,000; Nia approves higher purchases and changes to the promised D16 date. No further money or borrowing is approved.
17-B — Status reconciliation, close 1 October.
Status reconciliation, close 1 October. PM: “90% complete”:90/100 development tickets closed,30 of the closed tickets optional. The number of optional tickets among the ten still open is unrecorded. Import log 9,000 rows includes 1,200 duplicate rows. Finance’s signed register accepts 6,500 distinct records;1,300 more distinct records await resolution. Required total 10,000 distinct. A supplier’s “7,000 approved” includes 500 of those pending records that it has reviewed provisionally without finance acceptance. Critical scenarios:12 pass on the current candidate, eight fail, ten unrun; all 30 must undergo final UAT. Six of twenty recently closed defects reopened. Estimates below include known work; neither that reopening proportion nor ticket completion supplies a scenario probability.
17-C — Resource calendar, functional managers,4 October.
Resource calendar, functional managers,4 October. One engineer E and one data lead D are available every weekday. Test lead Q is available from D9. Finance SME F is available only D2,4,7,9,12,14,17,19,22,24,27,29,32,34, subject to the joint-state change below. Each is one full person-day: no concurrent assignments. F alone accepts finance logic/reconciliation. Optional rebates demand five E-days and dashboards three D-days, currently promised in W1. Unlisted parallel activity does not shorten the work.
17-D — Remaining core work, technical/finance leads,4 October.
Remaining core work, technical/finance leads,4 October.

| Work | Required effort and dependencies |
|---|---|
| A tax/credit corrections | Six E-days, or eight under J2. F also on the ACTUAL first and last E-day. E-days may have gaps. |
| B data resolution | Two consecutive D-days; then one D+F day; then two consecutive D-days; then one D+F day. Four ordered blocks, with gaps allowed. |
| C historical orders | Four D-days, may split; independent of A/B. |
| N native stock interface | Three consecutive E-days, independent of A/B/C. |
| X temporary alternative to N | Two consecutive E-days THEN one F-day. Validated alternative; adds four finance hours/week for eight weeks after launch, removing improvement capacity. No extra cash cost beyond specified burn. |
| L final conversion | Two consecutive E+D days after A/B/C and N or X complete; Q also on day 2. Normally starts Monday only. |
| U final UAT | Three ordered, separate Q+F days after L; rerun all 30 cases. No configuration/data changes during U. |
| R cutover/rollback rehearsal | Two consecutive E+D+Q days after U passes; F also on day 2. Includes operational training/readiness. |

Gate:10,000 distinct records accepted, all 30 FINAL scenarios passed, balanced rehearsal reconciliation, demonstrated rollback and operations’ signature. F owns finance evidence; Q owns tests; Imani recommends; Nia authorises cutover. Future bookings are not passed results.
17-E — Upfront offers, commercial lead,4 October.
Upfront offers, commercial lead,4 October. Offers must be selected before work starts D1 and are non-refundable, including unused reservations. No equivalent capacity/slots can be bought after the state is observed.

| Offer | Effect | Fixed price |
|---|---|---:|
| S | Second qualified engineer D4–D6; may perform N only | £6,000 |
| T | Second qualified data analyst D1–D2; may perform C only | £2,000 |
| F8/F11/F13/F15 | Release the SAME F on the named day; buy any subset | £3,000 EACH |
| O | One out-of-cycle L start at D9; other L rules unchanged | £12,000 |
| K | Reserve conditional backfill allowing F on D20,D21,D23,D25, ONLY following verified J3 native fault, reported by Q no later than end D17 | £10,000 total, paid even if never activated |

K’s four days are available once its condition is met; there is still only one F. K cannot be activated for A rework, ordinary acceleration, an unverified suspicion or a policy that uses X and never encounters the native fault. Its price includes the backfill; no extra activation charge. It provides no E/D/Q or conversion capacity.
17-F — Joint scenario model and information release, finance/QA,4 October.
Joint scenario model and information release, finance/QA,4 October. Use these three mutually exclusive joint states for the policy comparison; do not multiply marginal risks or invent a fourth combination. They are bounded planning scenarios, not a guarantee that every real future outcome is covered.

| State | Work and capacity consequences | When the policy can know |
|---|---|---|
| J1: clean recovery | A six E-days; base F calendar; no additional native-interface fault | End D6 establishes A=6 and F17 available, but does NOT distinguish J1 from J3. |
| J2: tax interpretation expands | A eight E-days; F17 withdrawn for a related audit; no native fault | Independent finance review at end D6 confirms both changes, regardless of how A has been scheduled. |
| J3: native fault | A six E-days; base F calendar; if N is used, a stock-allocation fault appears at the END of the second actual U day | Before that observation, its information is identical to J1. Using X prevents this native fault, but does not reveal J3 early or make K eligible. |

If J3 occurs with N, repair requires three consecutive E+D days AFTER detection, with F on the last actual day. Then restart all three U days: earlier U results do not carry. L does not repeat because converted data is unaffected. Repair is the complete additional work for this state. Alternatively, replace N with X using its stated E/F sequence, then restart U; its ongoing finance burden applies. No purchase, repair, switch or capacity release may be scheduled earlier than the information/authority permitting it. Later actions may legitimately differ once states are distinguishable.
17-G — Competing planning views,1 October; estimates, not measured frequencies.
Competing planning views,1 October; estimates, not measured frequencies. Delivery lead assigns J1/J2/J3 probabilities 0.55/0.25/0.20; finance lead uses 0.30/0.50/0.20. Neither view has authority to overrule the other. Do not silently average them. Sales expects a one-off £12,000 net trading contribution if all depots launch D16; finance confirms the calculation for that date, but its receipt is outside the project account and it cannot fund delivery or reduce gross authority. There is no additional contribution at D26/D36. No further loss/benefit estimate is supplied. Operations accepts supported D26/D36 as feasible operating choices if Nia approves.
17-H — Claims for review,4 October.
Claims for review,4 October. Supplier: “Our £92,000 accelerated plan is inside authority and makes 25 October possible.” Its quote includes base pre-cutover burn and acceleration purchases, not hypercare or any rework outcome. PM: “We can decide which extra people to buy once the tax review tells us more.” Finance analyst: “The £12,000 trading contribution can be subtracted from the funding request.” These are positions to assess, not amendments to 17-A–G. Nia has not chosen a risk preference between earlier trading, expected value, gross exposure and funding margin; make and defend a practical recommendation instead of refusing to choose.
Read the complete supplied case · BPB2-17
# BPB2-17 — Wholesale ERP implementation — recover delivery

Suite BPB2 v0.3-pilot; Foray 247, third essai.

### Scenario

Tern Wholesale’s ERP programme has missed two releases and promises all three depots a 25 October launch. At the start of 4 October 2027, programme director Imani Reed must purchase recovery capacity before learning whether tax work will expand or a native-interface fault will emerge. A trading opportunity rewards the earlier date, but project authority must also cover a credible fallback. Imani needs a sponsor recommendation that connects evidence, daily resource commitments, conditional decisions and affordable delivery across the supplied joint scenarios.

### Exact task prompt

Produce one usable in-browser answer page for the named practitioner, using the supplied fictional evidence. Keep observed facts, forecasts, assumptions and proposed decisions distinguishable, and cite source IDs beside material claims. No external research or real-world certification is required.

Recommend the recovery policy Nia should authorise now. Reconcile readiness and competing cost/value claims; stop optional work that obstructs the core. Compare an executable earliest-window policy with a lower-exposure policy, including upfront purchases, daily resource witnesses, observation-dependent decisions and a funded supported cutover in every supplied joint state. Show gross project cash, protected hypercare, contingent commercial value and remaining authority separately. Test both the delivery and finance probability views, or show the decision-switch threshold that covers them. State what can responsibly be promised, the evidence gates and reversal points. Decisions made before an observation must be identical in states that cannot yet be distinguished; do not assume perfect foresight, later purchases, an unsupported cutover or a cancelled project as the delivery fallback.

### Supplied records

**17-A — Sponsor mandate, Nia Holt, start 4 October 2027.** D1 is Monday 4 October; count weekdays, no holidays. A task finishes at its last day’s end; successors start next workday. Cutover is at the START of D16=25 October, D26=8 November or D36=22 November only. All three depots launch together; ordering, inventory, purchasing, tax and reconciliation remain core. Rebates/dashboards may be deferred; approved manual rebates may operate eight weeks. Choose upfront purchases before D1 assignments begin; their supplied D1 capacity is then usable. No state information is available at that decision.

Total remaining authority is £150,000, including an immutable £12,000 post-launch hypercare provision for any cutover date. Do not spend or net off that provision. Base programme burn is £4,000 per working day from D1 through the day before cutover, including idle days and all base people/vendors. Purchases are additional. Legacy operation through D36 is supported within this basis; no unpriced pause/discount is available. Imani may authorise aggregate recovery purchases≤£25,000; Nia approves higher purchases and changes to the promised D16 date. No further money or borrowing is approved.

**17-B — Status reconciliation, close 1 October.** PM: “90% complete”:90/100 development tickets closed,30 of the closed tickets optional. The number of optional tickets among the ten still open is unrecorded. Import log 9,000 rows includes 1,200 duplicate rows. Finance’s signed register accepts 6,500 distinct records;1,300 more distinct records await resolution. Required total 10,000 distinct. A supplier’s “7,000 approved” includes 500 of those pending records that it has reviewed provisionally without finance acceptance. Critical scenarios:12 pass on the current candidate, eight fail, ten unrun; all 30 must undergo final UAT. Six of twenty recently closed defects reopened. Estimates below include known work; neither that reopening proportion nor ticket completion supplies a scenario probability.

**17-C — Resource calendar, functional managers,4 October.** One engineer E and one data lead D are available every weekday. Test lead Q is available from D9. Finance SME F is available only D2,4,7,9,12,14,17,19,22,24,27,29,32,34, subject to the joint-state change below. Each is one full person-day: no concurrent assignments. F alone accepts finance logic/reconciliation. Optional rebates demand five E-days and dashboards three D-days, currently promised in W1. Unlisted parallel activity does not shorten the work.

**17-D — Remaining core work, technical/finance leads,4 October.**

| Work | Required effort and dependencies |
|---|---|
| A tax/credit corrections | Six E-days, or eight under J2. F also on the ACTUAL first and last E-day. E-days may have gaps. |
| B data resolution | Two consecutive D-days; then one D+F day; then two consecutive D-days; then one D+F day. Four ordered blocks, with gaps allowed. |
| C historical orders | Four D-days, may split; independent of A/B. |
| N native stock interface | Three consecutive E-days, independent of A/B/C. |
| X temporary alternative to N | Two consecutive E-days THEN one F-day. Validated alternative; adds four finance hours/week for eight weeks after launch, removing improvement capacity. No extra cash cost beyond specified burn. |
| L final conversion | Two consecutive E+D days after A/B/C and N or X complete; Q also on day 2. Normally starts Monday only. |
| U final UAT | Three ordered, separate Q+F days after L; rerun all 30 cases. No configuration/data changes during U. |
| R cutover/rollback rehearsal | Two consecutive E+D+Q days after U passes; F also on day 2. Includes operational training/readiness. |

Gate:10,000 distinct records accepted, all 30 FINAL scenarios passed, balanced rehearsal reconciliation, demonstrated rollback and operations’ signature. F owns finance evidence; Q owns tests; Imani recommends; Nia authorises cutover. Future bookings are not passed results.

**17-E — Upfront offers, commercial lead,4 October.** Offers must be selected before work starts D1 and are non-refundable, including unused reservations. No equivalent capacity/slots can be bought after the state is observed.

| Offer | Effect | Fixed price |
|---|---|---:|
| S | Second qualified engineer D4–D6; may perform N only | £6,000 |
| T | Second qualified data analyst D1–D2; may perform C only | £2,000 |
| F8/F11/F13/F15 | Release the SAME F on the named day; buy any subset | £3,000 EACH |
| O | One out-of-cycle L start at D9; other L rules unchanged | £12,000 |
| K | Reserve conditional backfill allowing F on D20,D21,D23,D25, ONLY following verified J3 native fault, reported by Q no later than end D17 | £10,000 total, paid even if never activated |

K’s four days are available once its condition is met; there is still only one F. K cannot be activated for A rework, ordinary acceleration, an unverified suspicion or a policy that uses X and never encounters the native fault. Its price includes the backfill; no extra activation charge. It provides no E/D/Q or conversion capacity.

**17-F — Joint scenario model and information release, finance/QA,4 October.** Use these three mutually exclusive joint states for the policy comparison; do not multiply marginal risks or invent a fourth combination. They are bounded planning scenarios, not a guarantee that every real future outcome is covered.

| State | Work and capacity consequences | When the policy can know |
|---|---|---|
| J1: clean recovery | A six E-days; base F calendar; no additional native-interface fault | End D6 establishes A=6 and F17 available, but does NOT distinguish J1 from J3. |
| J2: tax interpretation expands | A eight E-days; F17 withdrawn for a related audit; no native fault | Independent finance review at end D6 confirms both changes, regardless of how A has been scheduled. |
| J3: native fault | A six E-days; base F calendar; if N is used, a stock-allocation fault appears at the END of the second actual U day | Before that observation, its information is identical to J1. Using X prevents this native fault, but does not reveal J3 early or make K eligible. |

If J3 occurs with N, repair requires three consecutive E+D days AFTER detection, with F on the last actual day. Then restart all three U days: earlier U results do not carry. L does not repeat because converted data is unaffected. Repair is the complete additional work for this state. Alternatively, replace N with X using its stated E/F sequence, then restart U; its ongoing finance burden applies. No purchase, repair, switch or capacity release may be scheduled earlier than the information/authority permitting it. Later actions may legitimately differ once states are distinguishable.

**17-G — Competing planning views,1 October; estimates, not measured frequencies.** Delivery lead assigns J1/J2/J3 probabilities 0.55/0.25/0.20; finance lead uses 0.30/0.50/0.20. Neither view has authority to overrule the other. Do not silently average them. Sales expects a one-off £12,000 net trading contribution if all depots launch D16; finance confirms the calculation for that date, but its receipt is outside the project account and it cannot fund delivery or reduce gross authority. There is no additional contribution at D26/D36. No further loss/benefit estimate is supplied. Operations accepts supported D26/D36 as feasible operating choices if Nia approves.

**17-H — Claims for review,4 October.** Supplier: “Our £92,000 accelerated plan is inside authority and makes 25 October possible.” Its quote includes base pre-cutover burn and acceleration purchases, not hypercare or any rework outcome. PM: “We can decide which extra people to buy once the tax review tells us more.” Finance analyst: “The £12,000 trading contribution can be subtracted from the funding request.” These are positions to assess, not amendments to 17-A–G. Nia has not chosen a risk preference between earlier trading, expected value, gross exposure and funding margin; make and defend a practical recommendation instead of refusing to choose.

### Browser deliverable and completion conditions

A sponsor recovery brief with an upfront decision, observation/decision map and daily resource witnesses for each recommended policy's state branches. Required controls compare the two policies you actually recommend across the three supplied joint states and the declared probability views (or their probability threshold). Each selection must show its validated schedule, supported window, gross cost including protected hypercare, value and authority margin; preserve explicit infeasibility where applicable. Identical-history branches must share the same earlier decisions. Expected comparisons must change with the declared probability view, while gross funding gates remain state-based. A validator for arbitrary purchase combinations or a general schedule optimiser is optional, not required. Preserve original outputs and show failed/unfunded branches; a static risk matrix or retrospectively perfect schedule is insufficient.

The page must work in the supplied browser/runtime; a download button or file preview alone is insufficient. Include a clear “Return to gallery” link using the shared shell route, or ../index.html if no route is supplied. Required controls must change the dependent results, not just their appearance. Original evidence and the initial dated position must remain recoverable; hypothetical changes do not create approvals or alter source records. Optional exports are secondary.